CS344: Introduction to CS344: Introduction to Artificial Intelligence g Pushpak Bhattacharyya Pushpak Bhattacharyya CSE Dept., IIT Bombay IIT Bombay Lecture 20-21– Natural Language Parsing Parsing
Parsing of Sentences Parsing of Sentences
Are sentences flat linear structures? Why tree? � Is there a principle in branching � When should the constituent give rise � When should the constituent give rise to children? � What is the hierarchy building principle? What is the hierarchy building principle?
Structure Dependency: A Case Study I nterrogative I nversion I nterrogative I nversion � � (1) John will solve the problem. Will John solve the problem? Will John solve the problem? Declarative Interrogative (2) a. Susan must leave. (2) a. Susan must leave. Must Susan leave? Must Susan leave? b. Harry can swim. Can Harry swim? c c. Mary has read the book. Has Mary read the book? a y as ead e boo as a y ead e boo Bill is sleeping. Is Bill sleeping? d. ………………………………………………………. The section, “ Structure dependency a case study” here is adopted from a talk given by Howard Lasnik (2003) in Delhi university. g y ( ) y
Interrogative inversion Structure Independent (1 st attempt) e Independent (1 st attempt) St ct (3) I nterrogative inversion process Beginning with a declarative, invert the first and second words to construct an interrogative. Declarative Declarative Interrogative Interrogative (4) a. The woman must leave. * Woman the must leave? b. A sailor can swim. b. A sailor can swim. * Sailor a can swim? Sailor a can swim? c. No boy has read the book. * Boy no has read the book? d. My friend is sleeping. * Friend my is sleeping?
Interrogative inversion correct pairings correct pairings Compare the incorrect pairings in (4) with the � correct pairings in (5): t i i i (5) Declarative I nterrogative (5) a. The woman must leave. (5) Th t l M Must the woman leave? t th l ? b. A sailor can swim. Can a sailor swim? c No boy has read the book c. No boy has read the book. Has no boy read the book? Has no boy read the book? d. My friend is sleeping. Is my friend sleeping?
Interrogative inversion Structure Independent (2 nd attempt) p ( p ) (6) I nterrogative inversion process : � Beginning with a declarative, move the auxiliary B i i ith d l ti th ili verb to the front to construct an interrogative. Declarative Declarative Interrogative Interrogative (7) a. Bill could be sleeping. * Be Bill could sleeping? Could Bill be sleeping? Could Bill be sleeping? b. Mary has been reading. * Been Mary has reading? Has Mary been reading? y g c. Susan should have left. * Have Susan should left? Should Susan have left?
Structure independent (3 rd attempt): (8) Interrogative inversion process Beginning with a declarative, move the first B i i ith d l ti th fi t � auxiliary verb to the front to construct an interrogative. interrogative Declarative Interrogative (9) a. The man who is here can swim. * Is the man who here can swim? (9) Th h i h i * I th h h i ? b. The boy who will play has left. * Will the boy who play has left?
Structure Dependent Correct Pairings � For � For the the above above examples, examples, fronting fronting the the second second auxiliary verb gives the correct form: Declarative Declarative Interrogative Interrogative (10) a.The man who is here can swim. Can the man who is here swim? b.The boy who will play has left. Has the boy who will play left?
Natural transformations are structure dependent (11) Does the child acquiring English learn these properties? (11) Does the child acquiring English learn these properties? (12) We are not dealing with a peculiarity of English. No known human language has a transformational process g g p that would produce pairings like those in (4), (7) and (9), repeated below: (4) a. The woman must leave. * Woman the must leave? (7) a. Bill could be sleeping. * Be Bill could sleeping? (9) a The man who is here can swim * Is the man who here can swim? (9) a. The man who is here can swim. * Is the man who here can swim?
Deeper trees needed for capturing sentence structure This wont do! Flat structure! NP PP The AP PP book with the blue cover of poems big [ The big book of poems with the [ The big book of poems with the Blue cover ] is on the table.
Other languages English English NP PP The AP PP book with the blue cover big of poems NP Hi di Hindi AP PP PP kitaab kavita kii niil jilda vaalii badii [ niil jilda vaalii kavita kii kitaab ]
Other languages: contd English English NP PP The AP PP book with the blue cover big of poems NP B Bengali li AP PP PP ti bai kavitar niil malaat deovaa motaa [niil malaat deovaa kavitar bai ti]
PPs are at the same level: flat with respect to the head word “book” No distinction in terms of dominance or c command dominance or c-command NP PP The AP PP book with the blue cover of poems big [ The big book of poems with the [ The big book of poems with the Blue cover ] is on the table.
“Constituency test of Replacement” runs Constituency test of Replacement runs into problems � One-replacement: � I bought the big [book of poems with the � I bought the big [book of poems with the blue cover] not the small [one] � One-replacement targets book of poems One replacement targets book of poems with the blue cover � Another one-replacement: � Another one replacement: � I bought the big [book of poems] with the blue cover not the small [one] with the red blue cover not the small [one] with the red cover � One-replacement targets book of poems
More deeply embedded structure NP N’ 1 The AP N’ 2 N’ 3 big big 3 PP PP N N with the blue cover with the blue cover PP book of poems
To target N 1 ’ � I want [ NP this [ N’ big book of poems with the red cover] and not [ N that [ N one]] the red cover] and not [ N that [ N one]]
Bar-level projections � Add intermediate structures � Add intermediate structures � NP � (D) N’ � N � (AP) N | N (PP) | N (PP) N’ � (AP) N’ | N’ (PP) | N (PP) � () indicates optionality
New rules produce this tree NP N-bar N’ 1 The AP N’ 2 N’ 3 big big 3 PP PP N N with the blue cover with the blue cover PP book of poems
As opposed to this tree NP PP The AP PP book with the blue cover of poems big
V-bar � What is the element in verbs corresponding to one-replacement for corresponding to one replacement for nouns � do-so or did-so � do-so or did-so
As opposed to this tree NP PP The AP PP book with the blue cover of poems big
I [eat beans with a fork] VP PP eat NP with a fork beans No constituent that groups together V and NP and excludes No constituent that groups together V and NP and excludes PP
Need for intermediate constituents � I [eat beans] with a fork but Ram [does VP so] with a spoon so] with a spoon V 1 ’ VP � V’ V 2 ’ V’ � V’ (PP) V’ � V’ (PP) V’ � V (NP) PP V V NP with a fork eat beans
How to target V 1 ’ � I [eat beans with a fork], and Ram VP [does so] too. [does so] too. V 1 ’ VP � V’ V 2 ’ V’ � V’ (PP) V’ � V’ (PP) V’ � V (NP) PP V V NP with a fork eat beans
Parsing Algorithms Parsing Algorithms
A simplified grammar � S → NP VP � NP → DT N | N � VP → V ADV | V
A segment of English Grammar � S’ � (C) S � S � { NP/S’} VP � S � { NP/S} VP � VP � (AP+ ) (VAUX) V (AP+ ) ({ NP/S’} ) (AP+ ) (PP+ ) (AP+ ) ({ NP/S’} ) (AP+ ) (PP+ ) (AP+ ) � NP � (D) (AP+ ) N (PP+ ) � PP � P NP � AP � (AP) A ( )
Example Sentence Example Sentence People laugh eop e aug 2 3 1 These are positions Lexicon: Lexicon: This indicate that both People - N, V Noun and Verb is possible for the word Laugh Laugh - N, V N V “People” People
Top-Down Parsing Top-Down Parsing State Backup State Action ----------------------------------------------------------------------------------------------------- 1. ((S) 1) - - Position of input pointer 2. ((NP VP)1) - - 3a. ((DT N VP)1) ((N VP) 1) - 3b. ((N VP)1) - 3b ((N VP)1) - 4. ((VP)2) - Consume “People” 5a. ((V ADV)2) ((V)2) - 6. ((ADV)3) ((V)2) Consume “laugh” 5b. ((V)2) - - 6. ((.)3) - Consume “laugh” Termination Condition : All inputs over. No symbols remaining. Note: Input symbols can be pushed back.
Discussion for Top-Down Parsing � This kind of searching is goal driven � This kind of searching is goal driven. � Gives importance to textual precedence (rule precedence). � No regard for data, a priori (useless expansions made).
Recommend
More recommend
